Most convolutional neural network (CNN) models have various difficulties in identifying crop diseases owing to morphological and physiological changes in crop tissues, and cells. Furthermore, a single crop disease can show different symptoms. Usually, the differences in symptoms between early crop disease and late crop disease stages include the area of disease and color of disease. This also poses additional difficulties for CNN models. Here, we propose a lightweight CNN model called GrapeNet for the identification of different symptom stages for specific grape diseases. The main components of GrapeNet are residual blocks, residual feature fusion blocks (RFFBs), and convolution block attention modules. The residual blocks are used to deepen the network depth and extract rich features. To alleviate the CNN performance degradation associated with a large number of hidden layers, we designed an RFFB module based on the residual block. It fuses the average pooled feature map before the residual block input and the high-dimensional feature maps after the residual block output by a concatenation operation, thereby achieving feature fusion at different depths. In addition, the convolutional block attention module (CBAM) is introduced after each RFFB module to extract valid disease information. The obtained results show that the identification accuracy was determined as 82.99%, 84.01%, 82.74%, 84.77%, 80.96%, 82.74%, 80.96%, 83.76%, and 86.29% for GoogLeNet, Vgg16, ResNet34, DenseNet121, MobileNetV2, MobileNetV3_large, ShuffleNetV2_×1.0, EfficientNetV2_s, and GrapeNet. The GrapeNet model achieved the best classification performance when compared with other classical models. The total number of parameters of the GrapeNet model only included 2.15 million. Compared with DenseNet121, which has the highest accuracy among classical network models, the number of parameters of GrapeNet was reduced by 4.81 million, thereby reducing the training time of GrapeNet by about two times compared with that of DenseNet121. Moreover, the visualization results of Grad-cam indicate that the introduction of CBAM can emphasize disease information and suppress irrelevant information. The overall results suggest that the GrapeNet model is useful for the automatic identification of grape leaf diseases.

Laser speckle noise caused by coherence between lasers greatly influences the produced image. In order to suppress the effect of laser speckles on images, in this paper we set up a combination of a laser-structured light module and an infrared camera to acquire laser images, and propose an improved weighted non-local mean (IW-NLM) filtering method that adopts an SSI-based adaptive h-solving method to select the optimal h in the weight function. The analysis shows that the algorithm not only denoises the laser image but also smooths pixel jumps in the image, while preserving the image details. The experimental results show that compared with the original laser image, the equivalent number of looks (ENL) index of the IW-NLM filtered image improved by 0.80%. The speckle suppression index (SSI) of local images dropped from 4.69 to 2.55%. Compared with non-local mean filtering algorithms, the algorithm proposed in this paper is an improvement and provides more accurate data support for subsequent image processing analysis.

The utilization rate of ink liquid in the chamber is critical for the thermal bubble inkjet head. The difficult problem faced by the thermal bubble inkjet printing is how to maximize the use of ink in the chamber and increase the printing frequency. In this paper, by adding a flow restrictor and two narrow channels into the chamber, the H-shape flow-limiting structure is formed. At 1.8 μs, the speed of bubble expansion reaches the maximum, and after passing through the narrow channel, the maximum reverse flow rate of ink decreased by 25%. When the vapor bubble disappeared, the ink fills the nozzle slowly. At 20 μs, after passing through the narrow channel, the maximum flow rate of the ink increases by 39%. The inkjet printing frequency is 40 kHz, and the volume of the ink droplet is about 13.1 pL. The structure improves the frequency of thermal bubble inkjet printing and can maximize the use of liquid in the chamber, providing a reference for cell printing, 3D printing, bioprinting, and other fields.

It is crucial to improve printing frequency and ink droplet quality in thermal inkjet printing. This paper proposed a hemispherical chamber, and we used the CFD (computational fluid dynamics model) to simulate the inkjet process. During the whole simulation process, we first researched the hemispherical chamber’s inkjet state equipped with straight, conical shrinkage, and conical diffusion nozzles. Based on the broken time and volume of the liquid column, the nozzle geometry of the hemispherical chamber was determined to be a conical shrinkage nozzle with a specific size of 15 µm in height and 15 µm in diameter at the top, and 20 µm in diameter at the bottom. Next, we researched the inkjet performance of the square chamber, the round chamber, and the trapezoidal chamber. The round chamber showed the best inkjet performance using 1.8 µs as the driving time and 10 MPa as the maximum bubble pressure. After that, we compared the existing thermal inkjet printing heads. The results showed that the hemispherical chamber inkjet head had the best performance, achieving 30 KHz high-frequency printing and having the most significant volume ratio of droplet to the chamber, reaching 14.9%. As opposed to the current 15 KHz printing frequency of the thermal inkjet heads, the hemispherical chamber inkjet head has higher inkjet performance, and the volume ratio between the droplet and the chamber meets the range standard of 10–15%. The hemispherical chamber structure can be applied to thermal inkjet printing, office printing, 3D printing, and bio-printing.

Efficient printing frequency is critical for thermal bubble inkjet printing, while the difficulty lies in the structural design and material selection of the heating resistors. In this paper, a TaN film was used as the main material of the heating resistors, and two TaN films were placed in parallel to form the chopsticks-shaped structure. The heating time was divided into two sections, in which 0–0.1 μs was the preheating and 1.2–1.8 μs was the primary heating. At 1.8 μs, the maximum temperature of the Si₃N₄ film could reach about 1100 °C. At the same time, the SiO₂ film was added between the TaN film and Si₃N₄ film as a buffer layer, which effectively avoided the rupture of the Si₃N₄ film due to excessive thermal stress. Inside the inkjet print head, the maximum temperature of the chamber reached about 680 °C at 2.5 μs. Due to the high power of the heating resistors, the working time was greatly reduced and the frequency of the inkjet printing was effectively increased. At the interface between the back of the chip and the cartridge, the SiO₂ film was used to connect to ensure a timely ink supply. Under the condition of 12 V at 40 kHz, the inkjet chip could print efficiently with 10 nozzles at the same time. The inkjet chip proposed in this paper is not limited to only office printing, but also provides a new reference for 3D printing, cell printing, and vegetable and fruit printing.

The existing methods in multimodal sentiment analysis (MSA) primarily focus on designing sophisticated fusion networks to obtain multimodal representations. However, they may ignore some critical challenges such as interference from task-irrelevant background information and performance bottlenecks arising from inter-modal semantic conflict. To address these issues, this paper proposes a framework named Refined Separated Representation Learning and Conflict Alleviation Network (RSCAN), which aims to extract effective features for the task and fuse them in a conflict-aware way. Specifically, we design a two-layer Refined Separated Representation Learning Network (RSLN) that interactively separates effective features relevant to the prediction task from background redundancies for each modality. The process is reinforced by contrastive learning to alleviate inter-modal heterogeneity and promote feature alignment before and after separation. Additionally, dominated by text modality, the Conflict Alleviation Network (CAN) is designed to explicitly quantify cross-modal conflict and adaptively adjust the contributions of other modalities to generate high-quality fused representations for sentiment prediction, thereby enhancing the model’s robustness in conflicting scenarios. Extensive experiments on three benchmark datasets CMU-MOSI, CMU-MOSEI, and CH-SIMS demonstrate the effectiveness of our task-oriented feature refinement and conflict-aware fusion strategy in improving overall sentiment analysis performance.